Global color characterization is a very powerful tool to model in a simple yet discriminant way the visual appearance of complex objects. A fixed reference model of this type can ...
It is difficult to adapt discriminative classifiers, particularly kernel based ones such as support vector machines (SVMs), to handle mismatches between the training and test da...
This paper describes recent advances in hidden Markov model (HMM) based OCR for machine-printed Arabic documents. A combination of scriptindependent and script-specific techniques...
Abstract— This paper proposes an approach to learn subjectindependent P300 models for EEG-based brain-computer interfaces. The P300 models are first learned using a pool of exis...
This paper proposes a novel method of estimating 3-D hand posture from images observed in complex backgrounds. Conventional methods often cause mistakes by mis-matches of local ima...